{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "view-in-github",
"colab_type": "text"
},
"source": [
"![\"Open](\"https://colab.research.google.com/assets/colab-badge.svg\")
"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "NKpZ6xVn9Zbd"
},
"outputs": [],
"source": [
"import tensorflow as tf\n",
"from tensorflow.keras import datasets, layers, models\n",
"import matplotlib.pyplot as plt"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "oKj5jLg29n_m",
"outputId": "0dbbf627-7d06-4279-cda8-2d16ddfbee40"
},
"outputs": [
{
"data": {
"text/plain": [
""
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"tf.device('/device:GPU:0')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "7asdUr3HTJrb",
"outputId": "34055f50-3b26-462a-d97a-dc784fd40ae5"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Mounted at /content/drive\n"
]
}
],
"source": [
"from google.colab import drive\n",
"drive.mount('/content/drive')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "w17sptkG9q6x",
"outputId": "b99a581b-04ef-44fc-d737-1642dba82e8b"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 840 files belonging to 4 classes.\n",
"Found 360 files belonging to 4 classes.\n",
"['cataract', 'diabetic_retinopathy', 'glaucoma', 'normal']\n"
]
}
],
"source": [
"img_height = 128\n",
"img_width = 128\n",
"\n",
"train_dataset = tf.keras.utils.image_dataset_from_directory(\n",
" '/content/drive/MyDrive/deep_learning/UTS/dataset/latih',\n",
" seed=12831294,\n",
" image_size=(img_width, img_height),\n",
" batch_size=32\n",
")\n",
"\n",
"test_dataset = tf.keras.utils.image_dataset_from_directory(\n",
" '/content/drive/MyDrive/deep_learning/UTS/dataset/uji',\n",
" seed=4285938,\n",
" image_size=(img_width, img_height),\n",
" batch_size=32\n",
")\n",
"\n",
"classes = train_dataset.class_names\n",
"print(classes)\n",
"\n",
"AUTOTUNE = tf.data.AUTOTUNE\n",
"train_dataset = train_dataset.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)\n",
"test_dataset = test_dataset.cache().prefetch(buffer_size=AUTOTUNE)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "-lWu0hed91R9"
},
"outputs": [],
"source": [
"from tensorflow.keras import Sequential\n",
"from tensorflow.keras.layers import Rescaling, RandomFlip, RandomRotation, RandomZoom\n",
"from tensorflow.keras.layers import Conv2D, BatchNormalization, Activation\n",
"from tensorflow.keras.layers import GlobalAveragePooling2D, ZeroPadding2D\n",
"from tensorflow.keras.layers import MaxPooling2D, Flatten, Dense, Dropout\n",
"from tensorflow.keras import regularizers"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "zCsQeghp92oX",
"outputId": "c1ced081-e4cb-4b8b-ebf2-489d240c19d0"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Model: \"sequential_1\"\n",
"_________________________________________________________________\n",
" Layer (type) Output Shape Param # \n",
"=================================================================\n",
" sequential (Sequential) (None, 128, 128, 3) 0 \n",
" \n",
" rescaling (Rescaling) (None, 128, 128, 3) 0 \n",
" \n",
" conv2d (Conv2D) (None, 126, 126, 16) 448 \n",
" \n",
" batch_normalization (BatchN (None, 126, 126, 16) 64 \n",
" ormalization) \n",
" \n",
" activation (Activation) (None, 126, 126, 16) 0 \n",
" \n",
" max_pooling2d (MaxPooling2D (None, 63, 63, 16) 0 \n",
" ) \n",
" \n",
" conv2d_1 (Conv2D) (None, 61, 61, 32) 4640 \n",
" \n",
" batch_normalization_1 (Batc (None, 61, 61, 32) 128 \n",
" hNormalization) \n",
" \n",
" activation_1 (Activation) (None, 61, 61, 32) 0 \n",
" \n",
" max_pooling2d_1 (MaxPooling (None, 30, 30, 32) 0 \n",
" 2D) \n",
" \n",
" conv2d_2 (Conv2D) (None, 28, 28, 64) 18496 \n",
" \n",
" batch_normalization_2 (Batc (None, 28, 28, 64) 256 \n",
" hNormalization) \n",
" \n",
" activation_2 (Activation) (None, 28, 28, 64) 0 \n",
" \n",
" max_pooling2d_2 (MaxPooling (None, 14, 14, 64) 0 \n",
" 2D) \n",
" \n",
" conv2d_3 (Conv2D) (None, 12, 12, 128) 73856 \n",
" \n",
" batch_normalization_3 (Batc (None, 12, 12, 128) 512 \n",
" hNormalization) \n",
" \n",
" activation_3 (Activation) (None, 12, 12, 128) 0 \n",
" \n",
" conv2d_4 (Conv2D) (None, 10, 10, 128) 147584 \n",
" \n",
" batch_normalization_4 (Batc (None, 10, 10, 128) 512 \n",
" hNormalization) \n",
" \n",
" activation_4 (Activation) (None, 10, 10, 128) 0 \n",
" \n",
" conv2d_5 (Conv2D) (None, 8, 8, 128) 147584 \n",
" \n",
" batch_normalization_5 (Batc (None, 8, 8, 128) 512 \n",
" hNormalization) \n",
" \n",
" activation_5 (Activation) (None, 8, 8, 128) 0 \n",
" \n",
" global_average_pooling2d (G (None, 128) 0 \n",
" lobalAveragePooling2D) \n",
" \n",
" flatten (Flatten) (None, 128) 0 \n",
" \n",
" dense (Dense) (None, 512) 66048 \n",
" \n",
" dense_1 (Dense) (None, 4) 2052 \n",
" \n",
"=================================================================\n",
"Total params: 462,692\n",
"Trainable params: 461,700\n",
"Non-trainable params: 992\n",
"_________________________________________________________________\n"
]
}
],
"source": [
"augmentation = Sequential([\n",
" RandomFlip(\"horizontal\", input_shape=(img_width, img_height, 3)),\n",
" RandomRotation(0.1),\n",
" RandomZoom(0.1)\n",
"])\n",
"\n",
"#Pre-processing\n",
"model = Sequential([\n",
" augmentation,\n",
" Rescaling(1./255., input_shape=(img_width, img_height, 3)),\n",
"\n",
" #Model CNN \n",
" Conv2D(16, (3, 3)),\n",
" BatchNormalization(),\n",
" Activation('relu'),\n",
" MaxPooling2D((2, 2)),\n",
"\n",
" Conv2D(32, (3, 3)),\n",
" BatchNormalization(),\n",
" Activation('relu'),\n",
" MaxPooling2D((2, 2)),\n",
"\n",
" Conv2D(64, (3, 3)),\n",
" BatchNormalization(),\n",
" Activation('relu'),\n",
" MaxPooling2D((2, 2)),\n",
" #GlobalAveragePooling2D(),\n",
"\n",
" Conv2D(128, (3, 3)),\n",
" BatchNormalization(),\n",
" Activation('relu'),\n",
" Conv2D(128, (3, 3)),\n",
" BatchNormalization(),\n",
" Activation('relu'),\n",
" Conv2D(128, (3, 3)),\n",
" BatchNormalization(),\n",
" Activation('relu'),\n",
" GlobalAveragePooling2D(),\n",
" #Dropout(0.1),\n",
"\n",
" Flatten(),\n",
" Dense(512, activation='relu'),\n",
" Dense(len(classes))\n",
"])\n",
"\n",
"model.summary()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "Sx8aoW4-97eF"
},
"outputs": [],
"source": [
"from tensorflow.keras.losses import SparseCategoricalCrossentropy, CategoricalCrossentropy\n",
"from keras.callbacks import ModelCheckpoint\n",
"from tensorflow.keras.optimizers import Adam"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "4b7soc584N_T",
"outputId": "ed1db7af-ca7c-41c0-fb8b-ca9908332f80"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.5449 - accuracy: 0.7857\n",
"Epoch 1: val_loss improved from inf to 2.04535, saving model to /content/drive/MyDrive/deep_learning/UTS/model/ModelCheckpoint.epoch01-loss2.05-acc0.55.hdf5\n",
"27/27 [==============================] - 110s 1s/step - loss: 0.5449 - accuracy: 0.7857 - val_loss: 2.0454 - val_accuracy: 0.5472\n",
"Epoch 2/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.4311 - accuracy: 0.8190\n",
"Epoch 2: val_loss improved from 2.04535 to 0.82783, saving model to /content/drive/MyDrive/deep_learning/UTS/model/ModelCheckpoint.epoch02-loss0.83-acc0.70.hdf5\n",
"27/27 [==============================] - 26s 961ms/step - loss: 0.4311 - accuracy: 0.8190 - val_loss: 0.8278 - val_accuracy: 0.7028\n",
"Epoch 3/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.4275 - accuracy: 0.8298\n",
"Epoch 3: val_loss did not improve from 0.82783\n",
"27/27 [==============================] - 25s 923ms/step - loss: 0.4275 - accuracy: 0.8298 - val_loss: 1.1039 - val_accuracy: 0.6611\n",
"Epoch 4/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.3684 - accuracy: 0.8488\n",
"Epoch 4: val_loss did not improve from 0.82783\n",
"27/27 [==============================] - 26s 943ms/step - loss: 0.3684 - accuracy: 0.8488 - val_loss: 1.3378 - val_accuracy: 0.6306\n",
"Epoch 5/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.3903 - accuracy: 0.8381\n",
"Epoch 5: val_loss did not improve from 0.82783\n",
"27/27 [==============================] - 27s 998ms/step - loss: 0.3903 - accuracy: 0.8381 - val_loss: 0.8932 - val_accuracy: 0.6833\n",
"Epoch 6/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.3818 - accuracy: 0.8381\n",
"Epoch 6: val_loss did not improve from 0.82783\n",
"27/27 [==============================] - 27s 998ms/step - loss: 0.3818 - accuracy: 0.8381 - val_loss: 2.7329 - val_accuracy: 0.5611\n",
"Epoch 7/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.3806 - accuracy: 0.8595\n",
"Epoch 7: val_loss did not improve from 0.82783\n",
"27/27 [==============================] - 27s 998ms/step - loss: 0.3806 - accuracy: 0.8595 - val_loss: 0.9414 - val_accuracy: 0.6528\n",
"Epoch 8/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.3472 - accuracy: 0.8512\n",
"Epoch 8: val_loss did not improve from 0.82783\n",
"27/27 [==============================] - 26s 973ms/step - loss: 0.3472 - accuracy: 0.8512 - val_loss: 1.3746 - val_accuracy: 0.6278\n",
"Epoch 9/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.3378 - accuracy: 0.8619\n",
"Epoch 9: val_loss did not improve from 0.82783\n",
"27/27 [==============================] - 25s 920ms/step - loss: 0.3378 - accuracy: 0.8619 - val_loss: 2.6623 - val_accuracy: 0.5472\n",
"Epoch 10/10\n",
"27/27 [==============================] - ETA: 0s - loss: 0.3976 - accuracy: 0.8345\n",
"Epoch 10: val_loss did not improve from 0.82783\n",
"27/27 [==============================] - 27s 964ms/step - loss: 0.3976 - accuracy: 0.8345 - val_loss: 3.0843 - val_accuracy: 0.5361\n"
]
}
],
"source": [
"filepath = '/content/drive/MyDrive/deep_learning/UTS/model/ModelCheckpoint.epoch{epoch:02d}-loss{val_loss:.2f}-acc{val_accuracy:.2f}.hdf5'\n",
"checkpoint = ModelCheckpoint(filepath=filepath, \n",
" monitor='val_loss',\n",
" verbose=1, \n",
" save_best_only=True,\n",
" mode='min')\n",
"\n",
"model.compile(optimizer=Adam(),\n",
" loss=SparseCategoricalCrossentropy(from_logits=True),\n",
" metrics=['accuracy'])\n",
"\n",
"history = model.fit(train_dataset,\n",
" epochs=10,\n",
" callbacks=[checkpoint],\n",
" validation_data=test_dataset)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 843
},
"id": "GpETLvl54XaW",
"outputId": "af11432d-49c1-4906-a7f8-e3c3687e9fac"
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"plt.plot(history.history['accuracy'], label='Accuracy')\n",
"plt.plot(history.history['val_accuracy'], label='Val. Accuracy')\n",
"plt.legend(loc='upper left')\n",
"plt.show()\n",
"\n",
"plt.plot(history.history['loss'], label='Loss')\n",
"plt.plot(history.history['val_loss'], label='Val. Loss')\n",
"plt.legend(loc='upper left')\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "LQGFmvb54YeW"
},
"outputs": [],
"source": [
"from tensorflow import keras\n",
"import numpy as np\n",
"import cv2\n",
"import tensorflow as tf\n",
"import matplotlib.pyplot as plt\n",
"import glob\n",
"import random"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "3CFJN3fe4cnO",
"outputId": "70490131-04b5-4c1d-fc94-61bc388d0943"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 360 files belonging to 4 classes.\n"
]
}
],
"source": [
"classes = ['Cataract', \n",
" 'Diabetic_retinopathy', \n",
" 'Glaucoma',\n",
" 'Normal']\n",
"\n",
"test_dataset = tf.keras.utils.image_dataset_from_directory(\n",
" '/content/drive/MyDrive/deep_learning/UTS/dataset/uji',\n",
" seed=4285938,\n",
" image_size=(img_width, img_height),\n",
" batch_size=32\n",
")\n",
"\n",
"AUTOTUNE = tf.data.AUTOTUNE\n",
"test_dataset = test_dataset.cache().prefetch(buffer_size=AUTOTUNE)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "QkC1-L964gFL"
},
"outputs": [],
"source": [
"model = keras.models.load_model('/content/drive/MyDrive/deep_learning/UTS/model/ModelCheckpoint.epoch02-loss0.83-acc0.70.hdf5')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
},
"id": "2EbDnWho4lV8",
"outputId": "436aac93-cbef-4cc2-8681-a5598c651446"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"1/1 [==============================] - 0s 180ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 35ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 105ms/step\n",
"1/1 [==============================] - 0s 127ms/step\n",
"1/1 [==============================] - 0s 223ms/step\n",
"1/1 [==============================] - 0s 164ms/step\n",
"1/1 [==============================] - 0s 50ms/step\n",
"1/1 [==============================] - 0s 45ms/step\n",
"1/1 [==============================] - 0s 47ms/step\n",
"1/1 [==============================] - 0s 51ms/step\n",
"1/1 [==============================] - 0s 44ms/step\n",
"1/1 [==============================] - 0s 44ms/step\n",
"1/1 [==============================] - 0s 44ms/step\n",
"1/1 [==============================] - 0s 48ms/step\n",
"1/1 [==============================] - 0s 45ms/step\n",
"1/1 [==============================] - 0s 51ms/step\n",
"1/1 [==============================] - 0s 43ms/step\n",
"1/1 [==============================] - 0s 50ms/step\n",
"1/1 [==============================] - 0s 44ms/step\n",
"1/1 [==============================] - 0s 43ms/step\n",
"1/1 [==============================] - 0s 45ms/step\n",
"1/1 [==============================] - 0s 46ms/step\n",
"1/1 [==============================] - 0s 43ms/step\n",
"1/1 [==============================] - 0s 52ms/step\n",
"1/1 [==============================] - 0s 68ms/step\n",
"1/1 [==============================] - 0s 43ms/step\n",
"1/1 [==============================] - 0s 43ms/step\n",
"1/1 [==============================] - 0s 42ms/step\n",
"1/1 [==============================] - 0s 43ms/step\n",
"1/1 [==============================] - 0s 45ms/step\n",
"1/1 [==============================] - 0s 45ms/step\n",
"1/1 [==============================] - 0s 43ms/step\n",
"1/1 [==============================] - 0s 52ms/step\n",
"1/1 [==============================] - 0s 50ms/step\n",
"1/1 [==============================] - 0s 32ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 32ms/step\n",
"1/1 [==============================] - 0s 32ms/step\n",
"1/1 [==============================] - 0s 36ms/step\n",
"1/1 [==============================] - 0s 31ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 33ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 34ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 34ms/step\n",
"1/1 [==============================] - 0s 34ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 28ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 36ms/step\n",
"1/1 [==============================] - 0s 31ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 32ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 32ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 30ms/step\n",
"1/1 [==============================] - 0s 34ms/step\n",
"1/1 [==============================] - 0s 31ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n",
"1/1 [==============================] - 0s 29ms/step\n"
]
},
{
"output_type": "display_data",
"data": {
"text/plain": [
"